Release diff
Every release shows exactly which services, migrations, and environment variables changed, before anyone approves it.
Release automation
Ship on Friday.
Roll back in seconds.
Capstan turns your deploy scripts into one readable pipeline with gates, approvals, and an eleven-second rollback.
Trusted by engineering teams at
- Northgate
- Kestrel Labs
- Tidewater
- Halberd
- Pinemark
- Quarry Systems
What release teams see in the first month
- Deploys coordinated
- 4.8M
- Median rollback
- 11s
- Deploy frequency
- 3.1x
- Pipeline availability
- 99.98%
across 1,240 production services in the last year
from the first alert to production running the last good artifact
more releases per week by the end of month one
measured over the last four quarters, including incidents
Three moves, every release.
No rewrite. Capstan reads what you already run, then adds the parts every team ends up building by hand.
-
Connect
Point Capstan at your repo and your cloud account. It maps your existing shell targets and Makefiles into ordered stages.
-
Gate
Choose what has to pass before production moves: test suites, a migration dry run, an approval from the service owner.
-
Ship
Promote the release, watch the live diff roll out per region, and send production back to the last good artifact in one click.
Built for the parts of shipping that hurt.
On-call handoff
The next shift opens one page and sees what shipped, what is gated, and what is still rolling out.
Feature flags
Flags and deploys share one approval path, so nothing ships twice or slips out behind an old branch.
Audit trail
Every approval, override, and rollback is recorded with who did it and what they wrote down at the time.
Instant rollback
Eleven-second median. No git archaeology at 2am, and no rebuilding an artifact that already existed an hour ago.
We went from two deploys a week to forty. The pipeline stopped being a secret.
Pricing
Priced per service, not per seat.
Add every engineer you have. We only count the services you deploy.
Solo
$0forever
One service, unlimited deploys, and the full CLI.
- 1 production service
- Unlimited deploys and rollbacks
- Community support
Most teams start here
Team
$24per service / month
Everything in Solo, plus the controls a shared service needs.
- Unlimited services and seats
- Approvals, gates, and feature flags
- Slack and email approvals
- 90-day audit history
Enterprise
Custom
For regulated teams running their own infrastructure.
- SSO and SCIM provisioning
- Private runners in your VPC
- On-prem control plane
- Named solutions engineer
Every plan includes the CLI, the API, and rollback. No per-minute runner billing.
Questions we get on the first call.
Does Capstan replace our CI?
No. Capstan sits after CI. Your tests keep running wherever they run today; Capstan decides what happens once they pass and records the result against the release.
What if our deploy is a script someone wrote in 2019?
That is the normal starting point. Capstan reads existing shell and Makefile targets and maps them into stages, so the first pipeline does not require rewriting anything.
How does rollback actually work?
Capstan records the exact artifact each release points at. Rolling back re-points production at the last good artifact and replays any migrations that were marked reversible.
Can approvals stay in Slack?
Yes. Approvals, rejections, and rollback confirmations can all be handled from Slack, writing the same audit record as the web interface.
Where does Capstan run?
The control plane is hosted in the US or the EU. Deploy runners execute inside your own cloud account, so artifacts and secrets never leave your network.